Fluid Phase Equilibria 205 (2003) 163–170
Liquid–vapour equilibrium of {xBF3 + (1 − x)n-butane}
at 195.49 K
Lino M.B. Dias a,1 , Rui P. Bonifácio a , Eduardo J.M. Filipe a ,
Jorge C.G. Calado a,∗ , Clare McCabe b , George Jackson c
a
Centro de Quı́mica Estrutural, Instituto Superior Técnico, 1049-001 Lisboa, Portugal
Department of Chemical Engineering, Colorado School of Mines, Golden, CO 80401, USA
Department of Chemical Engineering, Imperial College of Science Technology and Medicine, London SW7 2BY, UK
b
c
Received 14 June 2002; accepted 4 October 2002
Abstract
The saturation vapour pressure of {x BF3 + (1−x)n-C4 H10 } has been measured at 195.49 K. The system shows
large positive deviations from Raoult’s law and liquid–liquid immiscibility over a wide composition range whose
limits have been estimated as 0.39±0.03 < x > 0.9785. The excess molar Gibbs energies (GEm ) have been calculated
as a function of composition from the vapour pressure data. For the hypothetical equimolar mixture GEm (x = 0.5)
= (703.9 ± 29) J mol−1 . The results were interpreted using the statistical associating fluid theory for potentials of
variable range (SAFT-VR), as well as compared with those of related systems, such as (BF3 + n-pentane) and (BF3
+ xenon).
© 2002 Elsevier Science B.V. All rights reserved.
Keywords: Phase equilibria; SAFT; Boron trifluoride; Alkanes; Xenon
1. Introduction
Boron trifluoride, BF3 , is widely used as a catalyst in a number of industrial processes, including the
polymerization of olefins. In such processes the reacting mixture is usually made up of large amounts of
BF3 , alkanes and alkenes, besides other additives. A detailed knowledge of the phase equilibria of all the
corresponding binary mixtures is crucial to the optimization of the whole process. Mixtures of alkanes
and alkenes have been extensively investigated, but the same is not true of mixtures of hydrocarbons
with boron trifluoride. In this paper we present a vapour–liquid equilibrium study of (BF3 + n-C4 H10 )
∗
Corresponding author. Tel.: +351-1-8419316; fax: +351-1-8464455.
E-mail address: [email protected] (J.C.G. Calado).
1
Present address: Bayer AG Process Development, ZT-TE Fluid Process Technology—Thermophysical Properties and Thermodynamics, Building B310, 51368 Leverkusen, Germany.
0378-3812/02/$ – see front matter © 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 3 7 8 - 3 8 1 2 ( 0 2 ) 0 0 2 7 6 - 5
164
L.M.B. Dias et al. / Fluid Phase Equilibria 205 (2003) 163–170
mixtures, a non-reactive system whose behaviour is important to the understanding of the polymerization
process. As far as we are aware, no other work has been done on (BF3 + n-C4 H10 ). Some results for the
solubility of BF3 in n-C5 H12 , at three temperatures, have been reported in the literature by Cade et al. [1];
however, the data cover only a limited range of pressures.
In (BF3 + n-C4 H10 ) both components have relatively simple molecular structures, with few internal
degrees of freedom, which makes them good candidates for the testing of statistical theories of liquids.
BF3 is a quadrupolar (non-polar), planar molecule whereas n-C4 H10 is a chain molecule. They are both
eminently suitable to quantitative description by theoretically-based equations of state, such as that
provided by the statistical associating fluid theory [2] (SAFT). This is an approach in which shape effects
and different types of molecular interactions can be separated and quantified, with each contribution
dependent on molecular parameters that have a clear physical meaning, such as molecular size or range
of the interaction. SAFT also has the ability to account simultaneously for non-sphericity and association
effects. For a recent review of SAFT and its applications, the reader is directed to the comprehensive
articles of Economou [3] and Muller and Gubbins [4].
Vapour–liquid equilibria measurements of (BF3 + n-C4 H10 ) mixtures were carried out at 195.49 K,
and the corresponding excess Gibbs energy (GEm ) calculated. As expected, the system is highly non-ideal,
with a wide liquid–liquid immiscibility range. This behaviour is similar to that found for mixtures of
alkanes and perfluoroalkanes.
The results were interpreted using a recent version of the SAFT approach, developed to treat potentials
of variable attractive range, (SAFT-VR) [5,6]. Although the theory is able to qualitatively describe the
overall pattern of the (BF3 + n-C4 H10 ) phase diagram, the quantitative agreement is poor. Another system
involving BF3 recently studied in our group is that of (BF3 + Xe) [7]. Since there is considerable evidence,
both experimental and theoretical, that xenon behaves much like an n-alkane when mixed with a variety
of other substances, including perfluorinated compounds [8,9], it was deemed useful, for comparative
purposes to apply the same theoretical approach to the (BF3 + Xe) system.
2. Experimental
The vapour pressure measurements were carried out isothermally in an all glass apparatus, using a
triple-point cryostat. Ammonia was selected as the cryogenic substance (triple-point temperature, T =
195.49 K). The apparatus and techniques have already been described in detail elsewhere [10]. The
mixtures were prepared condensing known amounts of each component into a calibrated pyknometer,
with the amount of substance calculated from pVT measurements. The pressures were measured using a
differential capacitance manometer (Datametrics 370, with 3.4 Pa resolution), which had been calibrated
against mercury manometers.
The components and cryostat substance, boron trifluoride from Matheson, n-butane from Linde and
ammonia from AirLiquide, had mass fractions 0.995, 0.9995 and 0.999, respectively. They were all further
purified by fractionation in a low temperature column. The final purities were checked by measuring the
vapour pressure at determined temperatures. At the working temperature of 195.49 K, the vapour pressure
of BF3 was found to be 397.80 ± 0.06 kPa, in good agreement with the literature value of 398.13 kPa [11].
Likewise, the vapour pressure of n-butane at the same temperature was 1.40 ± 0.01 kPa, to be compared
with 1.36 kPa [12]. Finally, the triple-point pressure of ammonia, was 6.093 ± 0.004 kPa, very similar to
the recommended value of 6.080 ± 0.003 kPa [13].
L.M.B. Dias et al. / Fluid Phase Equilibria 205 (2003) 163–170
165
The calculations of the amount of substance and of the excess molar Gibbs energy, GEm , need some ancillary data. The values of the second virial coefficients have already been given both for BF3 [7] and n-butane
[12]. The cross second virial coefficients (B12 ) were estimated using a corresponding states method [14].
The resulting values were B12 = −814.01 cm3 mol−1 at 195.49 K and B12 = −281.18 cm3 mol−1 at
298.15 K.
3. Results
The vapour pressures, p, of {x BF3 + (1−x)n-C4 H10 }liquid mixtures at 195.49 K as a function of
the liquid mole fraction of BF3 , x, are recorded in Table 1 and plotted in Fig. 1. The mixture shows
large positive deviations from Raoult’s law and the vapour pressure remains constant (average value of
404.03 ± 0.25 kPa) over much of the composition range, a clear sign of phase separation. The presence
of two liquid phases could also be confirmed visually. The limits of immiscibility were estimated to be
Table 1
Vapour pressures, p, and excess molar Gibbs energies, GEm , of {x BF3 + (1−x)n-C4 H10 }(l) at 195.49 K; x is the BF3 mole fraction
in the liquid phase; xo the overall mole fraction of BF3 ; y the BF3 mole fraction in the vapour phase; δp are the pressure residuals
0
xBF
3
xBF3
yBF3
p (kPa)
δp (kPa)
GEm (J mol−1 )
0.0000
–
–
–
–
–
–
0.4242
0.9482
0.9668
0.9785
1.0000
0.0000
0.0825
0.1798
0.2277
0.2748
0.3314
0.3561
–
–
–
–
1.0000
0.0000
0.9778
0.9916
0.9934
0.9952
0.9964
0.9968
–
–
–
–
1.0000
1.35
59.00
151.78
185.01
245.74
320.29
364.48
403.82
404.10
403.74
404.45
406.76
–
−4.5
9.6
−5.2
−0.8
−3.5
3.5
–
–
–
–
–
0
92.81
232.54
261.34
332.83
411.89
462.73
–
–
–
–
–
Fig. 1. Vapour pressure of {x BF3 + (1−x)n-C4 H10 }(l)={yBF3 + (1−y)n-C4 H10 }(g) plotted against x and against y at 195.49 K.
The lines are interpolations of the experimental results.
166
L.M.B. Dias et al. / Fluid Phase Equilibria 205 (2003) 163–170
Fig. 2. Excess molar Gibbs energy, GEm , of {x BF3 + (1−x)n-C4 H10 } at 195.49 K. The dashed line is the Redlich–Kister function.
(0.39 ± 0.03 < x > 0.9785). Six (out of ten) mixtures lie outside the immiscibility gap. The composition
of the vapour phase, y (the n-butane mole fraction) and the excess molar Gibbs energy, GEm , were evaluated
following Barker’s method [15], which minimizes the pressure residuals, δp = p − pcal. The values of δp,
also given in Table 1, are a good indication of the consistency of the experimental data. The excess molar
Gibbs energy of each mixture, GEm , was calculated at zero pressure and fitted to a Redlich–Kister-type
equation,
GEm
= x1 x2 A + B(x1 − x2 ) + C(x1 − x2 )2
RT
(1)
The fitting parameters A, B and C and their standard deviations are A = (1.732 ± 0.072), B =
(−2.12 ± 0.20) and C = (1.25 ± 0.37). The GEm values are plotted in Fig. 2 against the liquid mole
fraction, x, resulting in a very asymmetric curve. It should be noted, however, that the miscibility zone
(six mixtures) covers less than half of the composition range (x < 0.4). For the hypothetical equimolar
liquid mixture, GEm (x = 0.5) = (703.9 ± 29) J mol−1 .
4. Discussion
A quantitative interpretation of the results was attempted using the statistical associating fluid theory
for potentials of variable attractive range, SAFT-VR [5,6]. The main expressions of the SAFT-VR theory
for the square-well potential have been presented elsewhere, and the reader is directed to the original
references for full details [5,6].
The SAFT-VR approach describes molecules as chains of m tangent hard spherical segments with the
attractive interactions modeled by a potential of variable attractive range. Each segment is characterized by
three parameters, namely the hard-sphere diameter, σ, and the depth, , and width, λ, of the potential well.
For the n-alkanes, a simple empirical relation between the number of carbon atoms, n, in the alkyl chain
and the number of spherical segments, m, in the model chain has been proposed [16] m = 1+(n−1)/3. A
value of m = 2.00 was therefore used for n-butane. The remaining parameters, σ, ε and λ, are determined
by fitting the theoretical expressions to the experimental vapor pressure and orthobaric liquid density
L.M.B. Dias et al. / Fluid Phase Equilibria 205 (2003) 163–170
167
Table 2
Optimized square-well intermolecular potential parameters for boron trifluoride, n-butane, n-pentane and xenona
Substance
m
σ (nm)
ε/k (K)
λ
BF3
n-Butane
n-Pentane
Xenon
1.60
2.00
2.33
1.00
0.3412
0.3887
0.3931
0.3849
272.6
256.3
265.0
243.9
1.300
1.501
1.505
1.478
a
m is the number of spherical segments, λ the range parameter, σ the diameter of each segment, and ε/k the well depth.
data, over the entire liquid range. All these parameters have already been determined for n-butane. For
BF3 , m was treated as a floating parameter and calculated, alongside σ, ε and λ, from the fitting to the
orthobaric properties of the pure substance. All parameters are listed in Table 2.
The usefulness of the BF3 parameters, here calculated for the first time, can be assessed from a comparison between the calculated values of the vapour pressure and liquid densities and the experimental
data from the literature [11,17] (Fig. 3). From the figure we see that with this model for BF3 we are able
to accurately describe the phase behaviour of the pure component.
It is believed that BF3 , a typical Lewis acid, does not self-associate, but forms a number of adduct
compounds when mixed with a variety of substances (Lewis bases, such as ammonia, ethers and alcohols).
A possible description of this molecule by the SAFT approach could, therefore, include an association
site to account for those strong interactions. However, in the present case association is not possible,
because the n-butane molecules are devoid of basic character. Moreover, preliminary results obtained
by modeling the BF3 molecule with an additional attractive site, did not lead to any improvement in the
description of the phase equilibria of either the pure substance or the (BF3 + n-C4 H10 ) binary mixture.
As a consequence, the possibility of such an attractive site was ignored.
Phase equilibria studies in mixtures require the determination of a number of cross-parameters for the
unlike interaction. For both the size and range parameters, the arithmetic mean was used, i.e.
σ12 =
(σ11 + σ22 )
2
(2)
Fig. 3. (a) Vapour pressure curve for BF3 . (b) Vapour–liquid coexistence densities for BF3 . The continuous curves are the
SAFT-VR results obtained using parameters fitted simultaneously to vapour pressure and saturated liquid densities.
168
L.M.B. Dias et al. / Fluid Phase Equilibria 205 (2003) 163–170
which is known as the Lorentz combination rule, and
σ11 λ11 + σ22 λ22
λ12 =
σ11 + σ22
(3)
For the energy parameters, allowance was made for deviations from the geometric mean (the well-known
Berthelot combining rule)
εij = ξ(εii εjj )1/2
(4)
The SAFT-VR description of the vapour-liquid-liquid equilibrium for the (BF3 + n-C4 H10 ) system,
at 195.49 K, is represented in Fig. 4a. The best fit to the experimental data is obtained with a value of
ξ = 0.866. The theory is able to qualitatively predict the highly asymmetric phase diagram of this system.
In the literature there is a limited amount of experimental data for the (BF3 + n-pentane) system [1].
SAFT-VR should be able to model this system with the same value of ξ = 0.866 for the unlike interaction
as was used for (BF3 + n-butane) (Fig. 4b). It is surprising that in the case of n-pentane, the theory appears
to predict the experimental results much better than it does with n-butane. However, it should be observed
that the experimental data cover only a very limited range of the phase diagram and show considerable
scatter.
A useful comparison can be drawn between these results and those for other binary mixtures involving
perfluorocompounds and n-alkanes. The first perfluoroalkane is CF4 , a tetrahedral molecule whose first
non-zero multipole moment is the octopole. The (CF4 + n-C4 H10 ) system has been extensively investigated over the whole fluid range. It exhibits a rather complex phase behaviour in the transition from types
II and III, in the classification of van Konynenburg and Scott [18]. The phase diagram of this system, as
well as those for other systems involving perfluorinated substances such as C2 F6 [9], and SF6 [19], can be
well described by the SAFT-VR equation of state [20]. This leads us to believe that the poor performance
of SAFT-VR in what concerns the (BF3 + n-C4 H10 ) mixtures is due to an inadequate description of the
BF3 molecule. Unlike the perfluoroalkanes previously studied, BF3 has a quadrupole moment, a planar
geometry and a double bond delocalized over the whole molecule. The combined presence of all these
factors, none of which can be explicitly taken into account by the theory, could be the reason for the poor
quantitative agreement between theory and experiment in the case of the (BF3 + n-C4 H10 ) system. The
Fig. 4. Vapour–liquid equilibrium for: (a) {x BF3 + (1−x)n-C4 H10 } at 195.49 K; (b) (BF3 + n-C5 H12 ) at 322 K (䊉), 339 K
(䊏) and 366 K (䉱). The symbols represent the experimental results and the SAFT-VR predictions are shown by the continuous
curve.
L.M.B. Dias et al. / Fluid Phase Equilibria 205 (2003) 163–170
169
Fig. 5. Vapour–liquid equilibrium for (BF3 + xenon) at 182.34 K. The symbols represent the experimental results and the
SAFT-VR predictions are shown by the curves. Dashed line obtained with a ξ = 0.832 and solid line ξ = 0.8.
fact that BF3 , although being lighter than CF4 (one fluorine atom less), is less volatile than CF4 , implies
that the BF3 –BF3 interactions are stronger than the CF4 –CF4 interactions and provides further evidence
of the different nature of the intermolecular forces in BF3 when compared with a typical perfluorinated
molecule.
Another class of interesting mixtures, related to the one under discussion, is that generated by the
substitution of the hydrocarbon by xenon. This follows from a recent finding that liquid xenon displays a
behaviour that resembles that of the n-alkanes when mixed with a number of substances, including perfluorinated compounds. These findings have been corroborated by the theory in a number of recent papers on
the phase behaviour of binary systems involving xenon [8,9,12,21]. The calculated SAFT parameters for
xenon lie within the average value of those used to describe the n-alkanes; moreover, interaction parameters obtained for binary mixtures containing alkanes can be transferred to the corresponding mixtures
involving xenon. If the analogy holds, the value of ξ = 0.866 taken from the (BF3 + n-butane) system
should apply to the corresponding (BF3 + xenon) system. However, this assumption is not confirmed by
the theory (Fig. 5), and a much better fitting is obtained with a value of ξ = 0.80 for the deviation of the
cross-energy parameter from the geometric mean. Hence it seems that the transferability of molecular
parameters observed for other xenon and alkane binary mixtures appears to break down for systems
involving BF3 . Again, we believe that this is due to the inadequacy of the underlying model for BF3 . A
more realistic description of the BF3 molecule than that provided by this version of SAFT-VR is needed
if we are to obtain a quantitative account of the phase diagram.
List of symbols
B12
cross second virial coefficient
g
gas
GEm
excess molar Gibbs energy
l
liquid
m
number of spherical segments
p
pressure
170
δp
T
x0
x
y
L.M.B. Dias et al. / Fluid Phase Equilibria 205 (2003) 163–170
pressure residuals
temperature
overall mole fraction of BF3
liquid phase mole fraction of BF3
gas phase mole fraction of BF3
Greek letters
ε
square well depth
λ
square well width
ξ
binary interaction parameter
ρ
density
σ
hard core diameter
Acknowledgements
LMBD wishes to thank Fundação para a Ciência e a Tecnologia for financial support. We are also
grateful to BP Chemicals for their support. We should also like to thank Amparo Galindo for useful
discussions.
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
G.N. Cade, R.E. Dunn, H.J. Hepp, J. Am. Chem. Soc. 68 (1946) 2454–2455.
W.G. Chapman, G. Jacksone, K.E. Gubbins, Mol. Phys. 65 (1988) 1057–1079.
I.G. Economou, Ind. Eng. Chem. Res. 41 (2001) 953–962.
E.A. Muller, K.E. Gubbins, Ind. Eng. Chem. Res. 40 (2001) 2193–2211.
A. Gil-Villegas, A. Galindo, P.J. Whitehead, S.J. Mills, G. Jackson, A.N. Burgess, J. Chem. Phys. 106 (1997) 4168–4186.
A. Galindo, L.A. Davies, A. Gil-Villegas, G. Jackson, Mol. Phys. 93 (1998) 241–252.
J.C.G. Calado, E.J.M. Filipe, J. Chem. Soc., Faraday Trans. 92 (1996) 215–218.
E.J.M. Filipe, E.J.S. Gomes de Azevedo, L.F.G. Martins, V.A.M. Soares, J.C.G. Calado, C. McCabe, G. Jackson, J. Phys.
Chem. B. 104 (2000) 1315–1321.
C. McCabe, L.M.B. Dias, G. Jackson, E.J.M. Filipe, Phys. Chem. Chem. Phys. 3 (2001) 2852–2855.
J.C.G. Calado, E.J.S. Gomes de Azevedo, V.A.M. Soares, Chem. Eng. Commun. 5 (1980) 149–163.
H.S. Both, Carter, J. Phys. Chem. 36 (1932) 1359–1363.
E.J.M. Filipe, L.F.G. Martins, J.C.G. Calado, C. McCabe, G. Jackson, J. Phys. Chem. B. 104 (2000) 1322–1325.
L.A. K Staveley, L.Q. Lobo, J.C.G. Calado, Cryogenics 21 (1981) 131–144.
R.C. Reid, J.M. Prausnitz, B.E. Poling, The Properties of Gases and Liquids, 4th ed., McGraw-Hill, Singapore, 1988.
J.A. Barker, Aust. J. Chem. 6 (1953) 207–210.
A.L. Archer, M.D. Amos, G. Jackson, I.A. McLure, Int. J. Thermophys. 17 (1996) 201–211.
R.L. Rowley, W.V. Wilding, J.L. Oscarson, N.A. Zundel, T.L. Marshall, T.E. Daubert, R.P. Danner, DIPPR® Data
Compilation of Pure Compound Properties, Design Institute for Physical Properties, AIChE, New York, 2001.
P.H. van Konynenburg, R.L. Scott, Philos. Trans. R. Soc. London A 298 (1980) 495.
L.M.B. Dias, T. Cordeiro, C. McCabe, E.J.M. Filipe, J.C.G. Calado, J. Phys. Chem., submitted for publication.
C. McCabe, A. Galindo, A. Gil-Villegas, G. Jackson, J. Phys. Chem. B 102 (1998) 8060–8069.
E.J.M. Filipe, L.M.B. Dias, J.C.G. Calado, C. McCabe, G. Jackson, Phys. Chem. Chem. Phys. 4 (2002) 1618–1621.